Long Document Text Summarizer using the T5 Transformer Model

This project implements an abstractive text summarization system for long documents using the T5 (Text-To-Text Transfer Transformer) model. It is designed to handle lengthy documents such as research papers and articles, providing comprehensive and coherent summaries while preserving key information.

---

✨ Features

• ✅ Efficient processing of long documents (e.g., research papers, articles)
• ✅ Abstractive summarization capability
• ✅ Fine-tuned T5 transformer model (LD-T5)
• ✅ Higher ROUGE scores compared to baseline models
• ✅ Support for various document formats, including PDFs
• ✅ User-friendly interface built with Streamlit

---

🧠 Model Architecture

The project utilizes the T5-base model with an encoder-decoder structure:

• Encoder: Processes the input document and creates contextual representations
• Decoder: Generates the summary based on the encoded representations
• Multi-head Self-Attention: Enables the model to consider multiple perspectives and capture long-range dependencies

Screenshots

1. Homepage

2. Summary Generation

3. Generated Summary